[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [OPSEC] FYI draft-ietf-opsec-blackhole-urpf-04



(coffee != sleep) & (!coffee == sleep)
Donald.Smith at qwest.com gcia   

> -----Original Message-----
> From: opsec-bounces at ietf.org [mailto:opsec-bounces at ietf.org] 
> On Behalf Of Joel Jaeggli
> Sent: Monday, June 08, 2009 10:34 PM
> To: 'opsec at ietf.org'
> Subject: [OPSEC] FYI draft-ietf-opsec-blackhole-urpf-04
> 
> Another iteration of this draft after last call has been posted.
> 
> you many peruse it at your leisure. The diff located here:
> 
> http://tools.ietf.org/rfcdiff?difftype=--hwdiff&url2=draft-iet
> f-opsec-blackhole-urpf-04.txt
> 
> 
> shows the changes which are minor for the most part, except 
> for the very
> strong disclaimer now in 4.0...
> 
> Before enabling uRPF (in any mode), it is vital that you
>    fully understand the implications of doing so:
> 
>      - Strict mode will cause the router to drop all ingress traffic
>        if the best path back to the source address of the traffic is
>        not the interface from which the traffic was received.
>        Asymetric routing will cause strict mode uRPF to drop
>        legitimate traffic.
This completely ignores Junipers feasible-paths which allows for a urpf check to "pass" the packet if
a feasible path exists via the interface the packet arrived on.

We should request all vendors implement feasible-path and active-paths too since that would remove most of the danger in strict mode!

> 
>     - Loose mode causes the router to check if a route for the source
>       address of the traffic exists. This may also cause legitimate
>       traffic to be discarded.
> 
>    It is hoped that in the future, vendors will implement a "DoS-
>    mitigation" mode in addition to the Loose and Strict modes 
> -- in this
>    mode, the uRPF check will only fail if the next-hop for 
> the source of
>    the packet is a discard interface.
> _______________________________________________
> OPSEC mailing list
> OPSEC at ietf.org
> https://www.ietf.org/mailman/listinfo/opsec
>